SlideShare une entreprise Scribd logo
1  sur  5
Télécharger pour lire hors ligne
CONTINUOUSLY IMPROVE THE PERFORMANCE
OF PLANNING AND SCHEDULING MODELS WITH
PARAMETER FEEDBACK
Jeffrey D. Kelly* and Danielle Zyngier
Abstract
Continuously improving the accuracy and precision of planning and scheduling models is not new;
unfortunately it is not institutionalized in practice. The intent of this paper is to highlight a relatively
simple approach to historize or memorize past and present actual planning and scheduling data collected
into what we call the past rolling horizon (PRH). The PRH is identical to the future rolling horizon
(FRH) used in hierarchical production planning and model predictive control to manage omnipresent
uncertainty in the model and data. Instead of optimizing future decisions such as throughputs, operating-
modes and conditions we now optimize or estimate key model parameters. Although bias-updating using
a single time-sample of data is common practice in advanced process control and optimization to
incorporate “parameter” feedback, this is only realizable for real-time applications with comprehensive
measurement systems. Proposed in this paper is the use of multiple synchronous or asynchronous time-
samples in the past in conjunction with simultaneous reconciliation and regression to update a subset of
the model parameters on a past rolling horizon basis to improve the performance of planning and
scheduling models.
Keywords
Rolling horizons, reconciliation, parameter estimation, error-in-variables, closed-loop, feedback.
Introduction
* To whom all correspondence should be addressed. jdkelly@industrialgorithms.ca (Industrial Algorithms LLC.)
Planning and scheduling decision-making is
traditionally based on simplified models that can, with any
luck, accurately and precisely interpolate and extrapolate
the dominate behavior of the underlying production in
terms of its processes, operations and maintenance.
Related to this is the well known notion that "all models
are wrong, but some are useful" (G.E.P. Box) which
implies that even detailed models do not guarantee their
accuracy and precision in being able to predict and
optimize production (Forbes and Marlin, 1994, Zyngier
and Marlin, 2006).
The focus of this paper is to suggest a symmetrical
methodology to the future rolling horizon (FRH, Baker and
Peterson, 1979) we call the past rolling horizon (PRH)
which implements a continuous improvement strategy
similar to the Deming Wheel, Shewhart Cycle, Kaizen or
the Plan-Perform-Perfect-Loop (Kelly, 2005). The use of a
past rolling horizon for planning and scheduling is a
similar concept to moving-window estimators (Robertson
et al., 1996, Zyngier et al., 2001, Yip and Marlin, 2002). In
the PRH, “parameter” optimization is performed with
essentially fixed variables looking backwards in time
whereas in the FRH of planning and scheduling we use
“variable” optimization with fixed parameters looking
forwards in time where fixed implies exogenously defined
as opposed to endogenously determined via the
optimization process.
The structure of paper is to first illuminate the
different aspects of a model, second to highlight the issues
with typically passive data collection, third, updating and
estimating techniques are discussed and fourth a
motivating example is overviewed to demonstrate the need
for what we call “parameter” feedback and not just
“variable” feedback used in existing planning and
scheduling implementations.
Model Morphology
For decision-making found in the process industries,
models can usually be segregated into three different types.
(1) Macro or lumped models are the ones that usually
consider an entire plant (or a large section thereof) taking
into account only the core or critical unit-operations as
well as aggregations of units. Macro models are commonly
encountered in planning, scheduling and yield accounting
applications; (2) Micro or distributed models are more
restricted in scope than macro models in that they usually
consider a smaller number of unit-operations and are
usually distributed across several spatial dimensions
including time. The modeling of these unit-operations
comprise more detailed mass, energy and momentum
balances including vapor-liquid equilibrium and reaction
kinetic relationships. Micro models are widely applied in
advanced process simulation and optimization; and finally
(3) Molecular models address the relationships that occur
at an atomic or elemental level of granularity within a
small section of a unit-operation with very detailed
thermodynamics and transport phenomena.
Whether a model is macro, micro or molecular, there
are three relevant aspects of the model morphology. A
model may be classified by its structural form such as if it
is linear, piece-wise linear, polynomial, rational, multi-
linear or non-linear. On the other hand, its functional form
relates to its parameters, coefficients and/or factors. Its
syntactical form relates to how the model functions,
formulae or formulations are expressed. Syntactically,
models can be explicit or implicit (i.e., use closed- or open-
form1 equations respectively) of which the latter is the
more general form i.e., comprises explicit models as a
subset.
Other aspects of a model such as whether it is static or
dynamic (steady or unsteady), continuous or discrete and
deterministic or non-deterministic (stochastic or chaotic) is
also worth mentioning. Planning and scheduling models
are usually dynamic in the sense of having multiple time-
periods built-up from essentially static models, have a mix
of continuous and discrete variables to represent the
quantity, logic and quality dimensions and are mostly
deterministic. Another important aspect of a model is
related to its fidelity and size. Bigger and more detailed
models do not guarantee its precision or accuracy as shown
1 There can also be “pried-open” models which break-apart the
internal convergence loops inside closed-form models.
by Forbes and Marlin (1994), i.e., smaller and simpler
models can be just as “useful” if they meet certain point-
wise model accuracy criteria. Therefore, we can class
models into being either rigorous or rough. Rough models
are related to meta-models or surrogate models where a
blend of rigorous and rough sub-models is termed hybrid
modeling. Rigorous models are also known as first-
principle models and rough models are empirical models.
The types of models used in planning and scheduling are
mostly rough models where it is common practice to
linearize available rigorous models into first-order Taylor-
series expansions. These linearized models are called base
plus delta, fixed and variable, absolute and relative, slope
with intercept and shift models, (Bodington, 1995).
Data Issues
As is well known in the mathematical programming
community, any decision-making problem can be
decomposed into its model, data and solution. Therefore,
how to collect, clean and compile data for the purposes of
what we call “parameter” feedback merits some discussion.
As mentioned, the focus of this work is to establish a past
rolling horizon (PRH) for planning and scheduling
problems which is symmetrical to the future rolling
horizon (FRH) that exists at the heart of hierarchical
production planning (HPP) (Bitran and Hax, 1977) and
model predictive control (MPC) (Richalet et. al., 1978).
Ideally the data used to perform data reconciliation
and parameter estimation (DRPE), error-in-variables
method (EVM) (Reilly and Patino-Leal, 1981) or
instrumental variables regression (IVR) (Young, 1970)
should be independent and identically normally distributed,
else systemic or gross-errors in the data may exist hence
skewing the results. Unfortunately the data collected after a
plan or schedule has been completed is most often passive
and not perturbed, happenstance and not holistic and
degenerate and not designed. This means that the
calibration or training-data used to fit the key2 model
parameters in the PRH may not be representative of the
production or operating regions or ranges seen in the
control or testing-data found in the FRH. After all, the sole
purpose of planning and scheduling decision-making using
optimization is to push/pull the production to new and
more profitable/performant regimes perhaps not
implemented hitherto. Along this line, the quality of the
data can be classified into three main characteristics: (1)
diversity or richness of the data i.e., all sampled points
span different regions of the control-data, (2) consistency
of the data i.e., all sampled points in both the calibration-
and control-data are taken from the same system and (3)
statistical homogeneity of data i.e., all sampled points in
the calibration- and control-data have the same noise,
error, random shock/perturbation or uncertainty
2 See Krishnan et. al. (1992) or Zyngier (2006) to determine key
model parameters.
distributions including their non-linear correlation structure
(Rooney and Biegler, 2001).
To compound the issue, planning and scheduling also
forms a closed-loop feedback control circuit similar to that
found in MPC. The issues with structural analysis and
parameter estimation when closed-loop data is used were
first addressed by Box and MacGregor (1974) when fitting
linear and rational time-series transfer function models.
These issues also exist for planning and scheduling
models. Perhaps one of the main results of their work is to
introduce a small but persistent and uncorrelated dithering
signal or excitation to either the manipulated variables or
the set-points which continuously stimulates the process.
Additionally, closed-loop identification can be
implemented similarly to the approach of Koung and
MacGregor (1993). These same techniques can also be
applied to planning and scheduling optimization systems.
Finally, potential sources of error that exist in the data
arise from several diverse sources as enumerated by Kelly
(2000). They are forecast-errors, measurement-errors,
execution-errors (processing, operating and maintaining),
model structural- and functional-errors (including
decomposition- and aggregation-errors) and last but not
least, solution-errors due to the non-convexity of the
problems (existence of local optima).
Updating and Estimating Methods
The fundamental objective of any model updating and
estimating technique is to find the “best” functional form
which balances the tradeoffs between: (1) the best fit of the
calibration-data i.e., interpolation and (2) the most accurate
and precise parameter estimates when noise exists. There is
a third requirement which is the best prediction of the
control-data i.e., extrapolation, which is the overriding
goal of design-of-experiments and control-relevant
identification. Obviously the quality of the functional form
will depend on both the quality of the structural form and
the quality of the data discussed previously. And, in
advanced process or real-time optimization (RTO)
applications, recognition of the fidelity of the models must
be understood in order to increase the performance of the
models in terms of minimizing the offsets (accuracy)
between the true-plant’s response and the noise-free model
prediction and the variability (precision) of the predictions
due to disturbances (Yip and Marlin, 2004).
Although simple bias-updating is the standard
technique used in both MPC and RTO for “parameter”
feedback, it utilizes a single data point for one time-point
or period in the past and updates the bias, base, intercept or
fixed value of the model formula or equation only from the
measured “variable” feedback. Albeit this is sufficient to
asymptotically remove the offset between the actual and
assumed value of the plant, it is not particularly suited to
planning and scheduling systems. The reason is that in
MPC/RTO, real-time electronic and digitized
measurements of temperatures, pressures, flows, levels,
concentrations and properties are readily available.
Unfortunately, in planning and scheduling applications it
can take days, weeks or months to obtain measured
“variable” feedback given field/control laboratory,
accounting, billing and invoicing system delays.
Instead, a more sophisticated approach is necessary
which continuously collates data over the PRH and
performs a robust method of parameter estimation whilst
respecting errors in the variables. For our purposes, we
choose the method which incorporates simultaneously both
reconciliation and regression from Kelly (1998). This
method is identical to EVM but has useful diagnostics
tailored to the estimability and variability of the both the
reconciled and regressed estimates (Kelly and Zyngier,
2008). More specifically, it reliably computes the
observability of the parameters, the redundancy of the
measurements and the precision of the parameters and
adjusted measurements. In addition, it can also provide
necessary missing-data capability when some data sources
are not available usually intermittently.
Motivating Example
In order to illustrate the importance of “parameter”
feedback in addition to “variable” feedback, a simple
example is presented. In Figure 1, a system is shown that
consists of receiving a supply of material, processing it in a
reactor, storing it in a tank and shipping it to some demand
point. The reactor has a yield of product (Y) which is the
only uncertain parameter in this system. The demands are
exogenously defined by the customer, i.e., it is not a
degree-of-freedom when determining the plan or schedule.
Supply Reactor Tank Demand
True Plant
Tank Holdup
(Variable)
Supply
(Solution)
Reactor Yield
(Parameter)
Demand
(Parameter)
Figure. 1. Closed-Loop System.
At each planning and scheduling cycle, the supply
profile is dispatched to the true plant for implementation
after the solution is calculated. In terms of the feedback
mechanism, two strategies were compared: (i) “variable”
feedback only, i.e., inventory information is available at
the start of each cycle, and (ii) “variable” and “parameter”
feedback, i.e., both inventory and updated yield
information is available. For illustrative purposes, it is
assumed that there is no noise in the measurements and
therefore only regression is necessary to update Ymodel.
The equations used to determine the supply solution at
time-period t (St), given the demand (Dt) and assuming that
the inventory in the tank (It) must remain at a constant
target value (Itarget) of 2.0 is provided below.
tY/)DII(S elmodttetargtt  (1)
The inventory It in equation (1) is obtained through
“variable” feedback in that the value of the inventory at the
start of the cycle is measured and is used in the model for
the next cycle. The “true” inventory value is determined
after the supply profile is calculated by using the “true”
plant yield in equation (2):
tDYSII tplantttt   -1 (2)
Initially, Ymodel was assumed to be 0.7 whereas the true
plant yield Yplant has a value of 0.6. The results are shown
in Figure 2 with the demand profile the same for both
scenarios or situations. For the case where the yield is not
updated, i.e., there is “variable” feedback only, the dotted
line inventory profile shows an offset or bias from the
target inventory value of 2.0. By updating the yield at
every cycle using the PRH data, the offset from the
inventory target is quickly corrected (cycle 2) by the time
the Ymodel has been updated to the true value of 0.6.
1
2
3
4
5
6
7
0 1 2 3 4 5 6 7 8 9
Supply (true) Inventory (true)
Supply (fixed y) Inventory (fixed y)
Supply (updated y) Inventory (updated y)
Figure 2. Supply & Inventory Responses.
Therefore, it is evident that with “variable” feedback
only, it is impossible to remove the persistent offset or
inaccuracy in terms of meeting the planned/scheduled
inventory target of 2.0. Consequently, plan/schedule versus
actual reporting, common place in planning and scheduling
stewardship, will always display a non-zero bias when
significant parameter uncertainty exists of which “variable”
feedback alone will not correct.
Conclusions
Shown in this paper is the limitation of “variable”
feedback when moving from one planning and scheduling
cycle to another. Without both “variable” and “parameter”
feedback, offsets to planning and scheduling targets, set-
points and/or upper/lower bounds will exist similar to the
persistent offset found in proportional-only control and
those observed in real-time process optimization. By
employing reconciliation and regression technology on a
past rolling horizon (PRH) basis, it is possible to reduce
these offsets or inaccuracies asymptotically or evolutionary
over the life-time of the models.
References
Baker, K. R., Peterson, D. W. (1979). An Analytic Framework
for Evaluating Rolling Schedules.
Mgmt. Sci., 25, 341.
Bitran, G.R., Hax, A. C. (1977). One the Design of Hierarchical
Production Planning Systems. Decision Sciences, 8,
28.
Bodington, C.E. (1995), Planning, Scheduling and Control
Integration in the Process Industries, McGraw-Hill Inc.
Box, G.E.P. and MacGregor, J.F. (1974). The Analysis of Closed
Dynamic-Stochastic Systems, Technometrics, 16, 391.
Forbes, J. F., Marlin, T. E. (1994). Model Accuracy for
Economic Optimizing Controllers: The Bias Update
Case. Ind. Eng. Chem. Res., 33, 1919.
Kelly, J. D. (1998). A regularization approach to the
reconciliation of constrained data sets. Comp. Chem.
Eng., 22, 1771.
Kelly, J. D. (2000) The Necessity of Data Reconciliation: Some
Practical Issues. 2000 NPRA Computer Conference
Proceedings, Chicago, IL
Kelly, J. D. (2005). Modeling production-chain information.
Chem. Eng.. Prog. February, 28.
Kelly, J. D., Zyngier, D. (2008) A New and Improved MILP
Formulation to Optimize Observability, Redundancy
and Precision for Sensor Network Problems. AIChE J.,
doi:10.1002/aic.11475.
Koung, C.-W., MacGregor, J. F. (1993). Design of Identification
Experiments for Robust Control. A Geometric
Approach for Bivariate Processes. Ind. Eng. Chem.
Res., 32, 1658.
Krishnan, S., Barton, G.W. and Perkins, J.D. (1992). Robust
Parameter Estimation in On-Line Optimization – Part I.
Methodology and Simulated Case Study, Comp. chem..
Eng., 16, 6.
Reilly, P. M., Patino-Leal, H. (1981). A Bayesian Study of the
Error-in-Variables Model. Technometrics, 23, 221.
Richalet, J.A, Rault, J.L., Testud, and Papon, J., (1978). Model
Predictive Heuristic Control: Application to Industrial
Processes, Automatica, 14, 413.
Robertson, D. G., Lee, J. H., Rawlings, J. B. (1996). A Moving
Horizon-Based Approach for Least-Squares
Estimation. AIChE J., 42, 2209.
Rooney, W. C., Biegler, L. T. (2001). Design for Model
Parameter Uncertainty Using Nonlinear Confidence
Regions. AIChE J., 47, 1794.
Yip, W.S. and Marlin, T.E. (2002), Multiple Data Sets for Model
Updating in Real-Time Operations Optimization,
Comp. chem. Eng., 26, 1345.
Yip, W.S. and Marlin, T.E. (2004). The Effect of Model Fidelity
on Real-Time Optimization Performance, Comp. chem.
Eng., 28, 267.
Young, P.C. (1970). An Instrumental Variable Method for Real-
Time Identification of a Noisy Process, Automatica, 6,
271.
Zyngier, D., Araujo, O.Q.F., Coelho, M.A.Z., Lima, E.L. (2001)
Robust Soft Sensors fro SBR Monitoring, Water Sci.
Techn., 43, 101.
Zyngier, D. (2006) Monitoring, Diagnosing and Enhancing the
Performance of Linear Closed-Loop Real-Time
Optimization Systems, Ph.D. thesis, McMaster
University, Hamilton, ON, Canada.
Zyngier, D., Marlin, T.E. (2006) Monitoring and improving LP
optimization with uncertain parameters. In Proc. of
ESCAPE-16,Garmisch-Partenkirchen,Germany, 427.

Contenu connexe

Tendances

Integrating Fuzzy Dematel and SMAA-2 for Maintenance Expenses
Integrating Fuzzy Dematel and SMAA-2 for Maintenance ExpensesIntegrating Fuzzy Dematel and SMAA-2 for Maintenance Expenses
Integrating Fuzzy Dematel and SMAA-2 for Maintenance Expensesinventionjournals
 
83690136 sess-3-modelling-and-simulation
83690136 sess-3-modelling-and-simulation83690136 sess-3-modelling-and-simulation
83690136 sess-3-modelling-and-simulationnoogle1996
 
A NOVEL PERFORMANCE MEASURE FOR MACHINE LEARNING CLASSIFICATION
A NOVEL PERFORMANCE MEASURE FOR MACHINE LEARNING CLASSIFICATIONA NOVEL PERFORMANCE MEASURE FOR MACHINE LEARNING CLASSIFICATION
A NOVEL PERFORMANCE MEASURE FOR MACHINE LEARNING CLASSIFICATIONIJMIT JOURNAL
 
A Novel Performance Measure for Machine Learning Classification
A Novel Performance Measure for Machine Learning ClassificationA Novel Performance Measure for Machine Learning Classification
A Novel Performance Measure for Machine Learning ClassificationIJMIT JOURNAL
 
Influence over the Dimensionality Reduction and Clustering for Air Quality Me...
Influence over the Dimensionality Reduction and Clustering for Air Quality Me...Influence over the Dimensionality Reduction and Clustering for Air Quality Me...
Influence over the Dimensionality Reduction and Clustering for Air Quality Me...IJAEMSJORNAL
 
International Journal of Computational Engineering Research(IJCER)
International Journal of Computational Engineering Research(IJCER)International Journal of Computational Engineering Research(IJCER)
International Journal of Computational Engineering Research(IJCER)ijceronline
 
Churn Modeling For Mobile Telecommunications
Churn Modeling For Mobile TelecommunicationsChurn Modeling For Mobile Telecommunications
Churn Modeling For Mobile TelecommunicationsSalford Systems
 
FACE RECOGNITION USING PRINCIPAL COMPONENT ANALYSIS WITH MEDIAN FOR NORMALIZA...
FACE RECOGNITION USING PRINCIPAL COMPONENT ANALYSIS WITH MEDIAN FOR NORMALIZA...FACE RECOGNITION USING PRINCIPAL COMPONENT ANALYSIS WITH MEDIAN FOR NORMALIZA...
FACE RECOGNITION USING PRINCIPAL COMPONENT ANALYSIS WITH MEDIAN FOR NORMALIZA...csandit
 
TOPOLOGY AWARE LOAD BALANCING FOR GRIDS
TOPOLOGY AWARE LOAD BALANCING FOR GRIDS TOPOLOGY AWARE LOAD BALANCING FOR GRIDS
TOPOLOGY AWARE LOAD BALANCING FOR GRIDS ijgca
 
Reading week08 saaty_ahp_fundamentals
Reading week08 saaty_ahp_fundamentalsReading week08 saaty_ahp_fundamentals
Reading week08 saaty_ahp_fundamentalshenry KKK
 

Tendances (13)

Selfadaptive report
Selfadaptive reportSelfadaptive report
Selfadaptive report
 
Integrating Fuzzy Dematel and SMAA-2 for Maintenance Expenses
Integrating Fuzzy Dematel and SMAA-2 for Maintenance ExpensesIntegrating Fuzzy Dematel and SMAA-2 for Maintenance Expenses
Integrating Fuzzy Dematel and SMAA-2 for Maintenance Expenses
 
Basheka 244
Basheka 244Basheka 244
Basheka 244
 
83690136 sess-3-modelling-and-simulation
83690136 sess-3-modelling-and-simulation83690136 sess-3-modelling-and-simulation
83690136 sess-3-modelling-and-simulation
 
A NOVEL PERFORMANCE MEASURE FOR MACHINE LEARNING CLASSIFICATION
A NOVEL PERFORMANCE MEASURE FOR MACHINE LEARNING CLASSIFICATIONA NOVEL PERFORMANCE MEASURE FOR MACHINE LEARNING CLASSIFICATION
A NOVEL PERFORMANCE MEASURE FOR MACHINE LEARNING CLASSIFICATION
 
A Novel Performance Measure for Machine Learning Classification
A Novel Performance Measure for Machine Learning ClassificationA Novel Performance Measure for Machine Learning Classification
A Novel Performance Measure for Machine Learning Classification
 
Ey36927936
Ey36927936Ey36927936
Ey36927936
 
Influence over the Dimensionality Reduction and Clustering for Air Quality Me...
Influence over the Dimensionality Reduction and Clustering for Air Quality Me...Influence over the Dimensionality Reduction and Clustering for Air Quality Me...
Influence over the Dimensionality Reduction and Clustering for Air Quality Me...
 
International Journal of Computational Engineering Research(IJCER)
International Journal of Computational Engineering Research(IJCER)International Journal of Computational Engineering Research(IJCER)
International Journal of Computational Engineering Research(IJCER)
 
Churn Modeling For Mobile Telecommunications
Churn Modeling For Mobile TelecommunicationsChurn Modeling For Mobile Telecommunications
Churn Modeling For Mobile Telecommunications
 
FACE RECOGNITION USING PRINCIPAL COMPONENT ANALYSIS WITH MEDIAN FOR NORMALIZA...
FACE RECOGNITION USING PRINCIPAL COMPONENT ANALYSIS WITH MEDIAN FOR NORMALIZA...FACE RECOGNITION USING PRINCIPAL COMPONENT ANALYSIS WITH MEDIAN FOR NORMALIZA...
FACE RECOGNITION USING PRINCIPAL COMPONENT ANALYSIS WITH MEDIAN FOR NORMALIZA...
 
TOPOLOGY AWARE LOAD BALANCING FOR GRIDS
TOPOLOGY AWARE LOAD BALANCING FOR GRIDS TOPOLOGY AWARE LOAD BALANCING FOR GRIDS
TOPOLOGY AWARE LOAD BALANCING FOR GRIDS
 
Reading week08 saaty_ahp_fundamentals
Reading week08 saaty_ahp_fundamentalsReading week08 saaty_ahp_fundamentals
Reading week08 saaty_ahp_fundamentals
 

En vedette

Variable Selection Methods
Variable Selection MethodsVariable Selection Methods
Variable Selection Methodsjoycemi_la
 
Chicken Coop - the Sequel
Chicken Coop - the SequelChicken Coop - the Sequel
Chicken Coop - the Sequelophiesay
 
Techniques In Logo Design
Techniques In Logo DesignTechniques In Logo Design
Techniques In Logo Designslide32share
 
Formation chancellerie une gestion sensible au genre_mtimmerman_jan 2012
Formation chancellerie une gestion sensible au genre_mtimmerman_jan 2012Formation chancellerie une gestion sensible au genre_mtimmerman_jan 2012
Formation chancellerie une gestion sensible au genre_mtimmerman_jan 2012Elise Beyst
 
Advanced Production Accounting of an Olefins Plant Industrial Modeling Framew...
Advanced Production Accounting of an Olefins Plant Industrial Modeling Framew...Advanced Production Accounting of an Olefins Plant Industrial Modeling Framew...
Advanced Production Accounting of an Olefins Plant Industrial Modeling Framew...Alkis Vazacopoulos
 
P6 radioactive-decay
P6 radioactive-decayP6 radioactive-decay
P6 radioactive-decayopsonise
 
Сибирская кухня. блюда из дичи
Сибирская кухня. блюда из дичиСибирская кухня. блюда из дичи
Сибирская кухня. блюда из дичиMontikmur24
 
Grade8e test-120311084156-phpapp01 (1)
Grade8e test-120311084156-phpapp01 (1)Grade8e test-120311084156-phpapp01 (1)
Grade8e test-120311084156-phpapp01 (1)1shackealj
 
La santificacion del_creyente
La santificacion del_creyenteLa santificacion del_creyente
La santificacion del_creyenteTito Ortega
 
Презентація команди Васильківської ЗОШ І-ІІІ ступенів № 6
Презентація команди Васильківської ЗОШ І-ІІІ ступенів № 6Презентація команди Васильківської ЗОШ І-ІІІ ступенів № 6
Презентація команди Васильківської ЗОШ І-ІІІ ступенів № 608600 Vasilkov
 

En vedette (20)

Variable Selection Methods
Variable Selection MethodsVariable Selection Methods
Variable Selection Methods
 
Chicken Coop - the Sequel
Chicken Coop - the SequelChicken Coop - the Sequel
Chicken Coop - the Sequel
 
Mulleres artistas
Mulleres artistasMulleres artistas
Mulleres artistas
 
Jessica yousra
Jessica yousraJessica yousra
Jessica yousra
 
Techniques In Logo Design
Techniques In Logo DesignTechniques In Logo Design
Techniques In Logo Design
 
Hemais
HemaisHemais
Hemais
 
Formation chancellerie une gestion sensible au genre_mtimmerman_jan 2012
Formation chancellerie une gestion sensible au genre_mtimmerman_jan 2012Formation chancellerie une gestion sensible au genre_mtimmerman_jan 2012
Formation chancellerie une gestion sensible au genre_mtimmerman_jan 2012
 
Pendahuluan
PendahuluanPendahuluan
Pendahuluan
 
Ali Saruhan
Ali SaruhanAli Saruhan
Ali Saruhan
 
Advanced Production Accounting of an Olefins Plant Industrial Modeling Framew...
Advanced Production Accounting of an Olefins Plant Industrial Modeling Framew...Advanced Production Accounting of an Olefins Plant Industrial Modeling Framew...
Advanced Production Accounting of an Olefins Plant Industrial Modeling Framew...
 
6 vamos al mundo
6 vamos al mundo6 vamos al mundo
6 vamos al mundo
 
Chapter07
Chapter07Chapter07
Chapter07
 
P6 radioactive-decay
P6 radioactive-decayP6 radioactive-decay
P6 radioactive-decay
 
Introduktion til slideshare net
Introduktion til slideshare netIntroduktion til slideshare net
Introduktion til slideshare net
 
Сибирская кухня. блюда из дичи
Сибирская кухня. блюда из дичиСибирская кухня. блюда из дичи
Сибирская кухня. блюда из дичи
 
Grade8e test-120311084156-phpapp01 (1)
Grade8e test-120311084156-phpapp01 (1)Grade8e test-120311084156-phpapp01 (1)
Grade8e test-120311084156-phpapp01 (1)
 
La santificacion del_creyente
La santificacion del_creyenteLa santificacion del_creyente
La santificacion del_creyente
 
O día do libro
O día do libroO día do libro
O día do libro
 
Презентація команди Васильківської ЗОШ І-ІІІ ступенів № 6
Презентація команди Васильківської ЗОШ І-ІІІ ступенів № 6Презентація команди Васильківської ЗОШ І-ІІІ ступенів № 6
Презентація команди Васильківської ЗОШ І-ІІІ ступенів № 6
 
Portafolio
PortafolioPortafolio
Portafolio
 

Similaire à CONTINUOUSLY IMPROVE THE PERFORMANCE OF PLANNING AND SCHEDULING MODELS WITH PARAMETER FEEDBACK

Kelly zyngier oil&gasbookchapter_july2013
Kelly zyngier oil&gasbookchapter_july2013Kelly zyngier oil&gasbookchapter_july2013
Kelly zyngier oil&gasbookchapter_july2013Jeffrey Kelly
 
Unit-Operation Nonlinear Modeling for Planning and Scheduling Applications
Unit-Operation Nonlinear Modeling for Planning and Scheduling ApplicationsUnit-Operation Nonlinear Modeling for Planning and Scheduling Applications
Unit-Operation Nonlinear Modeling for Planning and Scheduling ApplicationsAlkis Vazacopoulos
 
Time series analysis
Time series analysisTime series analysis
Time series analysisFaltu Focat
 
Identification of repetitive processes at steady- and unsteady-state: Transfe...
Identification of repetitive processes at steady- and unsteady-state: Transfe...Identification of repetitive processes at steady- and unsteady-state: Transfe...
Identification of repetitive processes at steady- and unsteady-state: Transfe...Ricardo Magno Antunes
 
Generalized capital investment planning of oil-refineries using CPLEX-MILP an...
Generalized capital investment planning of oil-refineries using CPLEX-MILP an...Generalized capital investment planning of oil-refineries using CPLEX-MILP an...
Generalized capital investment planning of oil-refineries using CPLEX-MILP an...Alkis Vazacopoulos
 
Missing-Value Handling in Dynamic Model Estimation using IMPL
Missing-Value Handling in Dynamic Model Estimation using IMPL Missing-Value Handling in Dynamic Model Estimation using IMPL
Missing-Value Handling in Dynamic Model Estimation using IMPL Alkis Vazacopoulos
 
Assignment And Sequencing Models For Thescheduling Of Process Systems
Assignment And Sequencing Models For Thescheduling Of Process SystemsAssignment And Sequencing Models For Thescheduling Of Process Systems
Assignment And Sequencing Models For Thescheduling Of Process SystemsDustin Pytko
 
An Integrated Solver For Optimization Problems
An Integrated Solver For Optimization ProblemsAn Integrated Solver For Optimization Problems
An Integrated Solver For Optimization ProblemsMonica Waters
 
MOdelling CGE in GAMS
MOdelling CGE in GAMSMOdelling CGE in GAMS
MOdelling CGE in GAMSTarig Gibreel
 
Generalized capital investment planning of oil-refineries using MILP and sequ...
Generalized capital investment planning of oil-refineries using MILP and sequ...Generalized capital investment planning of oil-refineries using MILP and sequ...
Generalized capital investment planning of oil-refineries using MILP and sequ...optimizatiodirectdirect
 
Informing product design with analytical data
Informing product design with analytical dataInforming product design with analytical data
Informing product design with analytical dataTeam Consulting Ltd
 
COMPARISON BETWEEN THE GENETIC ALGORITHMS OPTIMIZATION AND PARTICLE SWARM OPT...
COMPARISON BETWEEN THE GENETIC ALGORITHMS OPTIMIZATION AND PARTICLE SWARM OPT...COMPARISON BETWEEN THE GENETIC ALGORITHMS OPTIMIZATION AND PARTICLE SWARM OPT...
COMPARISON BETWEEN THE GENETIC ALGORITHMS OPTIMIZATION AND PARTICLE SWARM OPT...IAEME Publication
 
Comparison between the genetic algorithms optimization and particle swarm opt...
Comparison between the genetic algorithms optimization and particle swarm opt...Comparison between the genetic algorithms optimization and particle swarm opt...
Comparison between the genetic algorithms optimization and particle swarm opt...IAEME Publication
 
analyzing-time-series-data-regression-with-a-practical-example.pptx
analyzing-time-series-data-regression-with-a-practical-example.pptxanalyzing-time-series-data-regression-with-a-practical-example.pptx
analyzing-time-series-data-regression-with-a-practical-example.pptxjoyadas092
 
analyzing-time-series-data-regression-with-a-practical-example (1).pptx
analyzing-time-series-data-regression-with-a-practical-example (1).pptxanalyzing-time-series-data-regression-with-a-practical-example (1).pptx
analyzing-time-series-data-regression-with-a-practical-example (1).pptxjoyadas092
 

Similaire à CONTINUOUSLY IMPROVE THE PERFORMANCE OF PLANNING AND SCHEDULING MODELS WITH PARAMETER FEEDBACK (20)

Kelly zyngier oil&gasbookchapter_july2013
Kelly zyngier oil&gasbookchapter_july2013Kelly zyngier oil&gasbookchapter_july2013
Kelly zyngier oil&gasbookchapter_july2013
 
Unit-Operation Nonlinear Modeling for Planning and Scheduling Applications
Unit-Operation Nonlinear Modeling for Planning and Scheduling ApplicationsUnit-Operation Nonlinear Modeling for Planning and Scheduling Applications
Unit-Operation Nonlinear Modeling for Planning and Scheduling Applications
 
Time series analysis
Time series analysisTime series analysis
Time series analysis
 
Identification of repetitive processes at steady- and unsteady-state: Transfe...
Identification of repetitive processes at steady- and unsteady-state: Transfe...Identification of repetitive processes at steady- and unsteady-state: Transfe...
Identification of repetitive processes at steady- and unsteady-state: Transfe...
 
Generalized capital investment planning of oil-refineries using CPLEX-MILP an...
Generalized capital investment planning of oil-refineries using CPLEX-MILP an...Generalized capital investment planning of oil-refineries using CPLEX-MILP an...
Generalized capital investment planning of oil-refineries using CPLEX-MILP an...
 
beven 2001.pdf
beven 2001.pdfbeven 2001.pdf
beven 2001.pdf
 
Missing-Value Handling in Dynamic Model Estimation using IMPL
Missing-Value Handling in Dynamic Model Estimation using IMPL Missing-Value Handling in Dynamic Model Estimation using IMPL
Missing-Value Handling in Dynamic Model Estimation using IMPL
 
IMPL Data Analysis
IMPL Data AnalysisIMPL Data Analysis
IMPL Data Analysis
 
StatsModelling
StatsModellingStatsModelling
StatsModelling
 
Assignment And Sequencing Models For Thescheduling Of Process Systems
Assignment And Sequencing Models For Thescheduling Of Process SystemsAssignment And Sequencing Models For Thescheduling Of Process Systems
Assignment And Sequencing Models For Thescheduling Of Process Systems
 
Shewhart
ShewhartShewhart
Shewhart
 
OR 14 15-unit_1
OR 14 15-unit_1OR 14 15-unit_1
OR 14 15-unit_1
 
An Integrated Solver For Optimization Problems
An Integrated Solver For Optimization ProblemsAn Integrated Solver For Optimization Problems
An Integrated Solver For Optimization Problems
 
MOdelling CGE in GAMS
MOdelling CGE in GAMSMOdelling CGE in GAMS
MOdelling CGE in GAMS
 
Generalized capital investment planning of oil-refineries using MILP and sequ...
Generalized capital investment planning of oil-refineries using MILP and sequ...Generalized capital investment planning of oil-refineries using MILP and sequ...
Generalized capital investment planning of oil-refineries using MILP and sequ...
 
Informing product design with analytical data
Informing product design with analytical dataInforming product design with analytical data
Informing product design with analytical data
 
COMPARISON BETWEEN THE GENETIC ALGORITHMS OPTIMIZATION AND PARTICLE SWARM OPT...
COMPARISON BETWEEN THE GENETIC ALGORITHMS OPTIMIZATION AND PARTICLE SWARM OPT...COMPARISON BETWEEN THE GENETIC ALGORITHMS OPTIMIZATION AND PARTICLE SWARM OPT...
COMPARISON BETWEEN THE GENETIC ALGORITHMS OPTIMIZATION AND PARTICLE SWARM OPT...
 
Comparison between the genetic algorithms optimization and particle swarm opt...
Comparison between the genetic algorithms optimization and particle swarm opt...Comparison between the genetic algorithms optimization and particle swarm opt...
Comparison between the genetic algorithms optimization and particle swarm opt...
 
analyzing-time-series-data-regression-with-a-practical-example.pptx
analyzing-time-series-data-regression-with-a-practical-example.pptxanalyzing-time-series-data-regression-with-a-practical-example.pptx
analyzing-time-series-data-regression-with-a-practical-example.pptx
 
analyzing-time-series-data-regression-with-a-practical-example (1).pptx
analyzing-time-series-data-regression-with-a-practical-example (1).pptxanalyzing-time-series-data-regression-with-a-practical-example (1).pptx
analyzing-time-series-data-regression-with-a-practical-example (1).pptx
 

Plus de Alkis Vazacopoulos

Automatic Fine-tuning Xpress-MP to Solve MIP
Automatic Fine-tuning Xpress-MP to Solve MIPAutomatic Fine-tuning Xpress-MP to Solve MIP
Automatic Fine-tuning Xpress-MP to Solve MIPAlkis Vazacopoulos
 
Amazing results with ODH|CPLEX
Amazing results with ODH|CPLEXAmazing results with ODH|CPLEX
Amazing results with ODH|CPLEXAlkis Vazacopoulos
 
Bia project poster fantasy football
Bia project poster  fantasy football Bia project poster  fantasy football
Bia project poster fantasy football Alkis Vazacopoulos
 
NFL Game schedule optimization
NFL Game schedule optimization NFL Game schedule optimization
NFL Game schedule optimization Alkis Vazacopoulos
 
2017 Business Intelligence & Analytics Corporate Event Stevens Institute of T...
2017 Business Intelligence & Analytics Corporate Event Stevens Institute of T...2017 Business Intelligence & Analytics Corporate Event Stevens Institute of T...
2017 Business Intelligence & Analytics Corporate Event Stevens Institute of T...Alkis Vazacopoulos
 
Very largeoptimizationparallel
Very largeoptimizationparallelVery largeoptimizationparallel
Very largeoptimizationparallelAlkis Vazacopoulos
 
Optimization Direct: Introduction and recent case studies
Optimization Direct: Introduction and recent case studiesOptimization Direct: Introduction and recent case studies
Optimization Direct: Introduction and recent case studiesAlkis Vazacopoulos
 
Informs 2016 Solving Planning and Scheduling Problems with CPLEX
Informs 2016 Solving Planning and Scheduling Problems with CPLEX Informs 2016 Solving Planning and Scheduling Problems with CPLEX
Informs 2016 Solving Planning and Scheduling Problems with CPLEX Alkis Vazacopoulos
 
Finite Impulse Response Estimation of Gas Furnace Data in IMPL Industrial Mod...
Finite Impulse Response Estimation of Gas Furnace Data in IMPL Industrial Mod...Finite Impulse Response Estimation of Gas Furnace Data in IMPL Industrial Mod...
Finite Impulse Response Estimation of Gas Furnace Data in IMPL Industrial Mod...Alkis Vazacopoulos
 
Industrial Modeling Service (IMS-IMPL)
Industrial Modeling Service (IMS-IMPL)Industrial Modeling Service (IMS-IMPL)
Industrial Modeling Service (IMS-IMPL)Alkis Vazacopoulos
 
Dither Signal Design Problem (DSDP) for Closed-Loop Estimation Industrial Mod...
Dither Signal Design Problem (DSDP) for Closed-Loop Estimation Industrial Mod...Dither Signal Design Problem (DSDP) for Closed-Loop Estimation Industrial Mod...
Dither Signal Design Problem (DSDP) for Closed-Loop Estimation Industrial Mod...Alkis Vazacopoulos
 
Distillation Curve Optimization Using Monotonic Interpolation
Distillation Curve Optimization Using Monotonic InterpolationDistillation Curve Optimization Using Monotonic Interpolation
Distillation Curve Optimization Using Monotonic InterpolationAlkis Vazacopoulos
 
Multi-Utility Scheduling Optimization (MUSO) Industrial Modeling Framework (M...
Multi-Utility Scheduling Optimization (MUSO) Industrial Modeling Framework (M...Multi-Utility Scheduling Optimization (MUSO) Industrial Modeling Framework (M...
Multi-Utility Scheduling Optimization (MUSO) Industrial Modeling Framework (M...Alkis Vazacopoulos
 
Advanced Parameter Estimation (APE) for Motor Gasoline Blending (MGB) Indust...
Advanced Parameter Estimation (APE) for Motor Gasoline Blending (MGB)  Indust...Advanced Parameter Estimation (APE) for Motor Gasoline Blending (MGB)  Indust...
Advanced Parameter Estimation (APE) for Motor Gasoline Blending (MGB) Indust...Alkis Vazacopoulos
 
Hybrid Dynamic Simulation (HDS) Industrial Modeling Framework (HDS-IMF)
Hybrid Dynamic Simulation (HDS)  Industrial Modeling Framework (HDS-IMF)Hybrid Dynamic Simulation (HDS)  Industrial Modeling Framework (HDS-IMF)
Hybrid Dynamic Simulation (HDS) Industrial Modeling Framework (HDS-IMF)Alkis Vazacopoulos
 

Plus de Alkis Vazacopoulos (20)

Automatic Fine-tuning Xpress-MP to Solve MIP
Automatic Fine-tuning Xpress-MP to Solve MIPAutomatic Fine-tuning Xpress-MP to Solve MIP
Automatic Fine-tuning Xpress-MP to Solve MIP
 
Data mining 2004
Data mining 2004Data mining 2004
Data mining 2004
 
Amazing results with ODH|CPLEX
Amazing results with ODH|CPLEXAmazing results with ODH|CPLEX
Amazing results with ODH|CPLEX
 
Bia project poster fantasy football
Bia project poster  fantasy football Bia project poster  fantasy football
Bia project poster fantasy football
 
NFL Game schedule optimization
NFL Game schedule optimization NFL Game schedule optimization
NFL Game schedule optimization
 
2017 Business Intelligence & Analytics Corporate Event Stevens Institute of T...
2017 Business Intelligence & Analytics Corporate Event Stevens Institute of T...2017 Business Intelligence & Analytics Corporate Event Stevens Institute of T...
2017 Business Intelligence & Analytics Corporate Event Stevens Institute of T...
 
Posters 2017
Posters 2017Posters 2017
Posters 2017
 
Very largeoptimizationparallel
Very largeoptimizationparallelVery largeoptimizationparallel
Very largeoptimizationparallel
 
Retail Pricing Optimization
Retail Pricing Optimization Retail Pricing Optimization
Retail Pricing Optimization
 
Optimization Direct: Introduction and recent case studies
Optimization Direct: Introduction and recent case studiesOptimization Direct: Introduction and recent case studies
Optimization Direct: Introduction and recent case studies
 
Informs 2016 Solving Planning and Scheduling Problems with CPLEX
Informs 2016 Solving Planning and Scheduling Problems with CPLEX Informs 2016 Solving Planning and Scheduling Problems with CPLEX
Informs 2016 Solving Planning and Scheduling Problems with CPLEX
 
ODHeuristics
ODHeuristicsODHeuristics
ODHeuristics
 
Finite Impulse Response Estimation of Gas Furnace Data in IMPL Industrial Mod...
Finite Impulse Response Estimation of Gas Furnace Data in IMPL Industrial Mod...Finite Impulse Response Estimation of Gas Furnace Data in IMPL Industrial Mod...
Finite Impulse Response Estimation of Gas Furnace Data in IMPL Industrial Mod...
 
Industrial Modeling Service (IMS-IMPL)
Industrial Modeling Service (IMS-IMPL)Industrial Modeling Service (IMS-IMPL)
Industrial Modeling Service (IMS-IMPL)
 
Dither Signal Design Problem (DSDP) for Closed-Loop Estimation Industrial Mod...
Dither Signal Design Problem (DSDP) for Closed-Loop Estimation Industrial Mod...Dither Signal Design Problem (DSDP) for Closed-Loop Estimation Industrial Mod...
Dither Signal Design Problem (DSDP) for Closed-Loop Estimation Industrial Mod...
 
Xmr im
Xmr imXmr im
Xmr im
 
Distillation Curve Optimization Using Monotonic Interpolation
Distillation Curve Optimization Using Monotonic InterpolationDistillation Curve Optimization Using Monotonic Interpolation
Distillation Curve Optimization Using Monotonic Interpolation
 
Multi-Utility Scheduling Optimization (MUSO) Industrial Modeling Framework (M...
Multi-Utility Scheduling Optimization (MUSO) Industrial Modeling Framework (M...Multi-Utility Scheduling Optimization (MUSO) Industrial Modeling Framework (M...
Multi-Utility Scheduling Optimization (MUSO) Industrial Modeling Framework (M...
 
Advanced Parameter Estimation (APE) for Motor Gasoline Blending (MGB) Indust...
Advanced Parameter Estimation (APE) for Motor Gasoline Blending (MGB)  Indust...Advanced Parameter Estimation (APE) for Motor Gasoline Blending (MGB)  Indust...
Advanced Parameter Estimation (APE) for Motor Gasoline Blending (MGB) Indust...
 
Hybrid Dynamic Simulation (HDS) Industrial Modeling Framework (HDS-IMF)
Hybrid Dynamic Simulation (HDS)  Industrial Modeling Framework (HDS-IMF)Hybrid Dynamic Simulation (HDS)  Industrial Modeling Framework (HDS-IMF)
Hybrid Dynamic Simulation (HDS) Industrial Modeling Framework (HDS-IMF)
 

Dernier

Transforming Data Streams with Kafka Connect: An Introduction to Single Messa...
Transforming Data Streams with Kafka Connect: An Introduction to Single Messa...Transforming Data Streams with Kafka Connect: An Introduction to Single Messa...
Transforming Data Streams with Kafka Connect: An Introduction to Single Messa...HostedbyConfluent
 
The Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptxThe Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptxMalak Abu Hammad
 
Presentation on how to chat with PDF using ChatGPT code interpreter
Presentation on how to chat with PDF using ChatGPT code interpreterPresentation on how to chat with PDF using ChatGPT code interpreter
Presentation on how to chat with PDF using ChatGPT code interpreternaman860154
 
SIEMENS: RAPUNZEL – A Tale About Knowledge Graph
SIEMENS: RAPUNZEL – A Tale About Knowledge GraphSIEMENS: RAPUNZEL – A Tale About Knowledge Graph
SIEMENS: RAPUNZEL – A Tale About Knowledge GraphNeo4j
 
AI as an Interface for Commercial Buildings
AI as an Interface for Commercial BuildingsAI as an Interface for Commercial Buildings
AI as an Interface for Commercial BuildingsMemoori
 
Swan(sea) Song – personal research during my six years at Swansea ... and bey...
Swan(sea) Song – personal research during my six years at Swansea ... and bey...Swan(sea) Song – personal research during my six years at Swansea ... and bey...
Swan(sea) Song – personal research during my six years at Swansea ... and bey...Alan Dix
 
From Event to Action: Accelerate Your Decision Making with Real-Time Automation
From Event to Action: Accelerate Your Decision Making with Real-Time AutomationFrom Event to Action: Accelerate Your Decision Making with Real-Time Automation
From Event to Action: Accelerate Your Decision Making with Real-Time AutomationSafe Software
 
GenCyber Cyber Security Day Presentation
GenCyber Cyber Security Day PresentationGenCyber Cyber Security Day Presentation
GenCyber Cyber Security Day PresentationMichael W. Hawkins
 
Factors to Consider When Choosing Accounts Payable Services Providers.pptx
Factors to Consider When Choosing Accounts Payable Services Providers.pptxFactors to Consider When Choosing Accounts Payable Services Providers.pptx
Factors to Consider When Choosing Accounts Payable Services Providers.pptxKatpro Technologies
 
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...shyamraj55
 
Salesforce Community Group Quito, Salesforce 101
Salesforce Community Group Quito, Salesforce 101Salesforce Community Group Quito, Salesforce 101
Salesforce Community Group Quito, Salesforce 101Paola De la Torre
 
08448380779 Call Girls In Friends Colony Women Seeking Men
08448380779 Call Girls In Friends Colony Women Seeking Men08448380779 Call Girls In Friends Colony Women Seeking Men
08448380779 Call Girls In Friends Colony Women Seeking MenDelhi Call girls
 
Key Features Of Token Development (1).pptx
Key  Features Of Token  Development (1).pptxKey  Features Of Token  Development (1).pptx
Key Features Of Token Development (1).pptxLBM Solutions
 
Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)Allon Mureinik
 
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
08448380779 Call Girls In Diplomatic Enclave Women Seeking MenDelhi Call girls
 
[2024]Digital Global Overview Report 2024 Meltwater.pdf
[2024]Digital Global Overview Report 2024 Meltwater.pdf[2024]Digital Global Overview Report 2024 Meltwater.pdf
[2024]Digital Global Overview Report 2024 Meltwater.pdfhans926745
 
Maximizing Board Effectiveness 2024 Webinar.pptx
Maximizing Board Effectiveness 2024 Webinar.pptxMaximizing Board Effectiveness 2024 Webinar.pptx
Maximizing Board Effectiveness 2024 Webinar.pptxOnBoard
 
Unblocking The Main Thread Solving ANRs and Frozen Frames
Unblocking The Main Thread Solving ANRs and Frozen FramesUnblocking The Main Thread Solving ANRs and Frozen Frames
Unblocking The Main Thread Solving ANRs and Frozen FramesSinan KOZAK
 
My Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 PresentationMy Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 PresentationRidwan Fadjar
 
Beyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
Beyond Boundaries: Leveraging No-Code Solutions for Industry InnovationBeyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
Beyond Boundaries: Leveraging No-Code Solutions for Industry InnovationSafe Software
 

Dernier (20)

Transforming Data Streams with Kafka Connect: An Introduction to Single Messa...
Transforming Data Streams with Kafka Connect: An Introduction to Single Messa...Transforming Data Streams with Kafka Connect: An Introduction to Single Messa...
Transforming Data Streams with Kafka Connect: An Introduction to Single Messa...
 
The Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptxThe Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptx
 
Presentation on how to chat with PDF using ChatGPT code interpreter
Presentation on how to chat with PDF using ChatGPT code interpreterPresentation on how to chat with PDF using ChatGPT code interpreter
Presentation on how to chat with PDF using ChatGPT code interpreter
 
SIEMENS: RAPUNZEL – A Tale About Knowledge Graph
SIEMENS: RAPUNZEL – A Tale About Knowledge GraphSIEMENS: RAPUNZEL – A Tale About Knowledge Graph
SIEMENS: RAPUNZEL – A Tale About Knowledge Graph
 
AI as an Interface for Commercial Buildings
AI as an Interface for Commercial BuildingsAI as an Interface for Commercial Buildings
AI as an Interface for Commercial Buildings
 
Swan(sea) Song – personal research during my six years at Swansea ... and bey...
Swan(sea) Song – personal research during my six years at Swansea ... and bey...Swan(sea) Song – personal research during my six years at Swansea ... and bey...
Swan(sea) Song – personal research during my six years at Swansea ... and bey...
 
From Event to Action: Accelerate Your Decision Making with Real-Time Automation
From Event to Action: Accelerate Your Decision Making with Real-Time AutomationFrom Event to Action: Accelerate Your Decision Making with Real-Time Automation
From Event to Action: Accelerate Your Decision Making with Real-Time Automation
 
GenCyber Cyber Security Day Presentation
GenCyber Cyber Security Day PresentationGenCyber Cyber Security Day Presentation
GenCyber Cyber Security Day Presentation
 
Factors to Consider When Choosing Accounts Payable Services Providers.pptx
Factors to Consider When Choosing Accounts Payable Services Providers.pptxFactors to Consider When Choosing Accounts Payable Services Providers.pptx
Factors to Consider When Choosing Accounts Payable Services Providers.pptx
 
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...
 
Salesforce Community Group Quito, Salesforce 101
Salesforce Community Group Quito, Salesforce 101Salesforce Community Group Quito, Salesforce 101
Salesforce Community Group Quito, Salesforce 101
 
08448380779 Call Girls In Friends Colony Women Seeking Men
08448380779 Call Girls In Friends Colony Women Seeking Men08448380779 Call Girls In Friends Colony Women Seeking Men
08448380779 Call Girls In Friends Colony Women Seeking Men
 
Key Features Of Token Development (1).pptx
Key  Features Of Token  Development (1).pptxKey  Features Of Token  Development (1).pptx
Key Features Of Token Development (1).pptx
 
Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)
 
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
 
[2024]Digital Global Overview Report 2024 Meltwater.pdf
[2024]Digital Global Overview Report 2024 Meltwater.pdf[2024]Digital Global Overview Report 2024 Meltwater.pdf
[2024]Digital Global Overview Report 2024 Meltwater.pdf
 
Maximizing Board Effectiveness 2024 Webinar.pptx
Maximizing Board Effectiveness 2024 Webinar.pptxMaximizing Board Effectiveness 2024 Webinar.pptx
Maximizing Board Effectiveness 2024 Webinar.pptx
 
Unblocking The Main Thread Solving ANRs and Frozen Frames
Unblocking The Main Thread Solving ANRs and Frozen FramesUnblocking The Main Thread Solving ANRs and Frozen Frames
Unblocking The Main Thread Solving ANRs and Frozen Frames
 
My Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 PresentationMy Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 Presentation
 
Beyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
Beyond Boundaries: Leveraging No-Code Solutions for Industry InnovationBeyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
Beyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
 

CONTINUOUSLY IMPROVE THE PERFORMANCE OF PLANNING AND SCHEDULING MODELS WITH PARAMETER FEEDBACK

  • 1. CONTINUOUSLY IMPROVE THE PERFORMANCE OF PLANNING AND SCHEDULING MODELS WITH PARAMETER FEEDBACK Jeffrey D. Kelly* and Danielle Zyngier Abstract Continuously improving the accuracy and precision of planning and scheduling models is not new; unfortunately it is not institutionalized in practice. The intent of this paper is to highlight a relatively simple approach to historize or memorize past and present actual planning and scheduling data collected into what we call the past rolling horizon (PRH). The PRH is identical to the future rolling horizon (FRH) used in hierarchical production planning and model predictive control to manage omnipresent uncertainty in the model and data. Instead of optimizing future decisions such as throughputs, operating- modes and conditions we now optimize or estimate key model parameters. Although bias-updating using a single time-sample of data is common practice in advanced process control and optimization to incorporate “parameter” feedback, this is only realizable for real-time applications with comprehensive measurement systems. Proposed in this paper is the use of multiple synchronous or asynchronous time- samples in the past in conjunction with simultaneous reconciliation and regression to update a subset of the model parameters on a past rolling horizon basis to improve the performance of planning and scheduling models. Keywords Rolling horizons, reconciliation, parameter estimation, error-in-variables, closed-loop, feedback. Introduction * To whom all correspondence should be addressed. jdkelly@industrialgorithms.ca (Industrial Algorithms LLC.) Planning and scheduling decision-making is traditionally based on simplified models that can, with any luck, accurately and precisely interpolate and extrapolate the dominate behavior of the underlying production in terms of its processes, operations and maintenance. Related to this is the well known notion that "all models are wrong, but some are useful" (G.E.P. Box) which implies that even detailed models do not guarantee their accuracy and precision in being able to predict and optimize production (Forbes and Marlin, 1994, Zyngier and Marlin, 2006). The focus of this paper is to suggest a symmetrical methodology to the future rolling horizon (FRH, Baker and Peterson, 1979) we call the past rolling horizon (PRH) which implements a continuous improvement strategy similar to the Deming Wheel, Shewhart Cycle, Kaizen or the Plan-Perform-Perfect-Loop (Kelly, 2005). The use of a past rolling horizon for planning and scheduling is a similar concept to moving-window estimators (Robertson et al., 1996, Zyngier et al., 2001, Yip and Marlin, 2002). In the PRH, “parameter” optimization is performed with essentially fixed variables looking backwards in time whereas in the FRH of planning and scheduling we use “variable” optimization with fixed parameters looking forwards in time where fixed implies exogenously defined as opposed to endogenously determined via the optimization process.
  • 2. The structure of paper is to first illuminate the different aspects of a model, second to highlight the issues with typically passive data collection, third, updating and estimating techniques are discussed and fourth a motivating example is overviewed to demonstrate the need for what we call “parameter” feedback and not just “variable” feedback used in existing planning and scheduling implementations. Model Morphology For decision-making found in the process industries, models can usually be segregated into three different types. (1) Macro or lumped models are the ones that usually consider an entire plant (or a large section thereof) taking into account only the core or critical unit-operations as well as aggregations of units. Macro models are commonly encountered in planning, scheduling and yield accounting applications; (2) Micro or distributed models are more restricted in scope than macro models in that they usually consider a smaller number of unit-operations and are usually distributed across several spatial dimensions including time. The modeling of these unit-operations comprise more detailed mass, energy and momentum balances including vapor-liquid equilibrium and reaction kinetic relationships. Micro models are widely applied in advanced process simulation and optimization; and finally (3) Molecular models address the relationships that occur at an atomic or elemental level of granularity within a small section of a unit-operation with very detailed thermodynamics and transport phenomena. Whether a model is macro, micro or molecular, there are three relevant aspects of the model morphology. A model may be classified by its structural form such as if it is linear, piece-wise linear, polynomial, rational, multi- linear or non-linear. On the other hand, its functional form relates to its parameters, coefficients and/or factors. Its syntactical form relates to how the model functions, formulae or formulations are expressed. Syntactically, models can be explicit or implicit (i.e., use closed- or open- form1 equations respectively) of which the latter is the more general form i.e., comprises explicit models as a subset. Other aspects of a model such as whether it is static or dynamic (steady or unsteady), continuous or discrete and deterministic or non-deterministic (stochastic or chaotic) is also worth mentioning. Planning and scheduling models are usually dynamic in the sense of having multiple time- periods built-up from essentially static models, have a mix of continuous and discrete variables to represent the quantity, logic and quality dimensions and are mostly deterministic. Another important aspect of a model is related to its fidelity and size. Bigger and more detailed models do not guarantee its precision or accuracy as shown 1 There can also be “pried-open” models which break-apart the internal convergence loops inside closed-form models. by Forbes and Marlin (1994), i.e., smaller and simpler models can be just as “useful” if they meet certain point- wise model accuracy criteria. Therefore, we can class models into being either rigorous or rough. Rough models are related to meta-models or surrogate models where a blend of rigorous and rough sub-models is termed hybrid modeling. Rigorous models are also known as first- principle models and rough models are empirical models. The types of models used in planning and scheduling are mostly rough models where it is common practice to linearize available rigorous models into first-order Taylor- series expansions. These linearized models are called base plus delta, fixed and variable, absolute and relative, slope with intercept and shift models, (Bodington, 1995). Data Issues As is well known in the mathematical programming community, any decision-making problem can be decomposed into its model, data and solution. Therefore, how to collect, clean and compile data for the purposes of what we call “parameter” feedback merits some discussion. As mentioned, the focus of this work is to establish a past rolling horizon (PRH) for planning and scheduling problems which is symmetrical to the future rolling horizon (FRH) that exists at the heart of hierarchical production planning (HPP) (Bitran and Hax, 1977) and model predictive control (MPC) (Richalet et. al., 1978). Ideally the data used to perform data reconciliation and parameter estimation (DRPE), error-in-variables method (EVM) (Reilly and Patino-Leal, 1981) or instrumental variables regression (IVR) (Young, 1970) should be independent and identically normally distributed, else systemic or gross-errors in the data may exist hence skewing the results. Unfortunately the data collected after a plan or schedule has been completed is most often passive and not perturbed, happenstance and not holistic and degenerate and not designed. This means that the calibration or training-data used to fit the key2 model parameters in the PRH may not be representative of the production or operating regions or ranges seen in the control or testing-data found in the FRH. After all, the sole purpose of planning and scheduling decision-making using optimization is to push/pull the production to new and more profitable/performant regimes perhaps not implemented hitherto. Along this line, the quality of the data can be classified into three main characteristics: (1) diversity or richness of the data i.e., all sampled points span different regions of the control-data, (2) consistency of the data i.e., all sampled points in both the calibration- and control-data are taken from the same system and (3) statistical homogeneity of data i.e., all sampled points in the calibration- and control-data have the same noise, error, random shock/perturbation or uncertainty 2 See Krishnan et. al. (1992) or Zyngier (2006) to determine key model parameters.
  • 3. distributions including their non-linear correlation structure (Rooney and Biegler, 2001). To compound the issue, planning and scheduling also forms a closed-loop feedback control circuit similar to that found in MPC. The issues with structural analysis and parameter estimation when closed-loop data is used were first addressed by Box and MacGregor (1974) when fitting linear and rational time-series transfer function models. These issues also exist for planning and scheduling models. Perhaps one of the main results of their work is to introduce a small but persistent and uncorrelated dithering signal or excitation to either the manipulated variables or the set-points which continuously stimulates the process. Additionally, closed-loop identification can be implemented similarly to the approach of Koung and MacGregor (1993). These same techniques can also be applied to planning and scheduling optimization systems. Finally, potential sources of error that exist in the data arise from several diverse sources as enumerated by Kelly (2000). They are forecast-errors, measurement-errors, execution-errors (processing, operating and maintaining), model structural- and functional-errors (including decomposition- and aggregation-errors) and last but not least, solution-errors due to the non-convexity of the problems (existence of local optima). Updating and Estimating Methods The fundamental objective of any model updating and estimating technique is to find the “best” functional form which balances the tradeoffs between: (1) the best fit of the calibration-data i.e., interpolation and (2) the most accurate and precise parameter estimates when noise exists. There is a third requirement which is the best prediction of the control-data i.e., extrapolation, which is the overriding goal of design-of-experiments and control-relevant identification. Obviously the quality of the functional form will depend on both the quality of the structural form and the quality of the data discussed previously. And, in advanced process or real-time optimization (RTO) applications, recognition of the fidelity of the models must be understood in order to increase the performance of the models in terms of minimizing the offsets (accuracy) between the true-plant’s response and the noise-free model prediction and the variability (precision) of the predictions due to disturbances (Yip and Marlin, 2004). Although simple bias-updating is the standard technique used in both MPC and RTO for “parameter” feedback, it utilizes a single data point for one time-point or period in the past and updates the bias, base, intercept or fixed value of the model formula or equation only from the measured “variable” feedback. Albeit this is sufficient to asymptotically remove the offset between the actual and assumed value of the plant, it is not particularly suited to planning and scheduling systems. The reason is that in MPC/RTO, real-time electronic and digitized measurements of temperatures, pressures, flows, levels, concentrations and properties are readily available. Unfortunately, in planning and scheduling applications it can take days, weeks or months to obtain measured “variable” feedback given field/control laboratory, accounting, billing and invoicing system delays. Instead, a more sophisticated approach is necessary which continuously collates data over the PRH and performs a robust method of parameter estimation whilst respecting errors in the variables. For our purposes, we choose the method which incorporates simultaneously both reconciliation and regression from Kelly (1998). This method is identical to EVM but has useful diagnostics tailored to the estimability and variability of the both the reconciled and regressed estimates (Kelly and Zyngier, 2008). More specifically, it reliably computes the observability of the parameters, the redundancy of the measurements and the precision of the parameters and adjusted measurements. In addition, it can also provide necessary missing-data capability when some data sources are not available usually intermittently. Motivating Example In order to illustrate the importance of “parameter” feedback in addition to “variable” feedback, a simple example is presented. In Figure 1, a system is shown that consists of receiving a supply of material, processing it in a reactor, storing it in a tank and shipping it to some demand point. The reactor has a yield of product (Y) which is the only uncertain parameter in this system. The demands are exogenously defined by the customer, i.e., it is not a degree-of-freedom when determining the plan or schedule. Supply Reactor Tank Demand True Plant Tank Holdup (Variable) Supply (Solution) Reactor Yield (Parameter) Demand (Parameter) Figure. 1. Closed-Loop System. At each planning and scheduling cycle, the supply profile is dispatched to the true plant for implementation after the solution is calculated. In terms of the feedback mechanism, two strategies were compared: (i) “variable” feedback only, i.e., inventory information is available at the start of each cycle, and (ii) “variable” and “parameter” feedback, i.e., both inventory and updated yield information is available. For illustrative purposes, it is assumed that there is no noise in the measurements and therefore only regression is necessary to update Ymodel. The equations used to determine the supply solution at time-period t (St), given the demand (Dt) and assuming that
  • 4. the inventory in the tank (It) must remain at a constant target value (Itarget) of 2.0 is provided below. tY/)DII(S elmodttetargtt  (1) The inventory It in equation (1) is obtained through “variable” feedback in that the value of the inventory at the start of the cycle is measured and is used in the model for the next cycle. The “true” inventory value is determined after the supply profile is calculated by using the “true” plant yield in equation (2): tDYSII tplantttt   -1 (2) Initially, Ymodel was assumed to be 0.7 whereas the true plant yield Yplant has a value of 0.6. The results are shown in Figure 2 with the demand profile the same for both scenarios or situations. For the case where the yield is not updated, i.e., there is “variable” feedback only, the dotted line inventory profile shows an offset or bias from the target inventory value of 2.0. By updating the yield at every cycle using the PRH data, the offset from the inventory target is quickly corrected (cycle 2) by the time the Ymodel has been updated to the true value of 0.6. 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 8 9 Supply (true) Inventory (true) Supply (fixed y) Inventory (fixed y) Supply (updated y) Inventory (updated y) Figure 2. Supply & Inventory Responses. Therefore, it is evident that with “variable” feedback only, it is impossible to remove the persistent offset or inaccuracy in terms of meeting the planned/scheduled inventory target of 2.0. Consequently, plan/schedule versus actual reporting, common place in planning and scheduling stewardship, will always display a non-zero bias when significant parameter uncertainty exists of which “variable” feedback alone will not correct. Conclusions Shown in this paper is the limitation of “variable” feedback when moving from one planning and scheduling cycle to another. Without both “variable” and “parameter” feedback, offsets to planning and scheduling targets, set- points and/or upper/lower bounds will exist similar to the persistent offset found in proportional-only control and those observed in real-time process optimization. By employing reconciliation and regression technology on a past rolling horizon (PRH) basis, it is possible to reduce these offsets or inaccuracies asymptotically or evolutionary over the life-time of the models. References Baker, K. R., Peterson, D. W. (1979). An Analytic Framework for Evaluating Rolling Schedules. Mgmt. Sci., 25, 341. Bitran, G.R., Hax, A. C. (1977). One the Design of Hierarchical Production Planning Systems. Decision Sciences, 8, 28. Bodington, C.E. (1995), Planning, Scheduling and Control Integration in the Process Industries, McGraw-Hill Inc. Box, G.E.P. and MacGregor, J.F. (1974). The Analysis of Closed Dynamic-Stochastic Systems, Technometrics, 16, 391. Forbes, J. F., Marlin, T. E. (1994). Model Accuracy for Economic Optimizing Controllers: The Bias Update Case. Ind. Eng. Chem. Res., 33, 1919. Kelly, J. D. (1998). A regularization approach to the reconciliation of constrained data sets. Comp. Chem. Eng., 22, 1771. Kelly, J. D. (2000) The Necessity of Data Reconciliation: Some Practical Issues. 2000 NPRA Computer Conference Proceedings, Chicago, IL Kelly, J. D. (2005). Modeling production-chain information. Chem. Eng.. Prog. February, 28. Kelly, J. D., Zyngier, D. (2008) A New and Improved MILP Formulation to Optimize Observability, Redundancy and Precision for Sensor Network Problems. AIChE J., doi:10.1002/aic.11475. Koung, C.-W., MacGregor, J. F. (1993). Design of Identification Experiments for Robust Control. A Geometric Approach for Bivariate Processes. Ind. Eng. Chem. Res., 32, 1658. Krishnan, S., Barton, G.W. and Perkins, J.D. (1992). Robust Parameter Estimation in On-Line Optimization – Part I. Methodology and Simulated Case Study, Comp. chem.. Eng., 16, 6. Reilly, P. M., Patino-Leal, H. (1981). A Bayesian Study of the Error-in-Variables Model. Technometrics, 23, 221. Richalet, J.A, Rault, J.L., Testud, and Papon, J., (1978). Model Predictive Heuristic Control: Application to Industrial Processes, Automatica, 14, 413. Robertson, D. G., Lee, J. H., Rawlings, J. B. (1996). A Moving Horizon-Based Approach for Least-Squares Estimation. AIChE J., 42, 2209. Rooney, W. C., Biegler, L. T. (2001). Design for Model Parameter Uncertainty Using Nonlinear Confidence Regions. AIChE J., 47, 1794. Yip, W.S. and Marlin, T.E. (2002), Multiple Data Sets for Model Updating in Real-Time Operations Optimization, Comp. chem. Eng., 26, 1345. Yip, W.S. and Marlin, T.E. (2004). The Effect of Model Fidelity on Real-Time Optimization Performance, Comp. chem. Eng., 28, 267. Young, P.C. (1970). An Instrumental Variable Method for Real- Time Identification of a Noisy Process, Automatica, 6, 271. Zyngier, D., Araujo, O.Q.F., Coelho, M.A.Z., Lima, E.L. (2001) Robust Soft Sensors fro SBR Monitoring, Water Sci. Techn., 43, 101. Zyngier, D. (2006) Monitoring, Diagnosing and Enhancing the Performance of Linear Closed-Loop Real-Time Optimization Systems, Ph.D. thesis, McMaster University, Hamilton, ON, Canada.
  • 5. Zyngier, D., Marlin, T.E. (2006) Monitoring and improving LP optimization with uncertain parameters. In Proc. of ESCAPE-16,Garmisch-Partenkirchen,Germany, 427.